Satish Sharma

TRAFFIC VOLUME TIME-SERIES ANALYSIS ACCORDING TO THE TYPE OF ROAD USE

Problems related to highway traffic operation and congestion management can be alleviated with the use of modern intelligent transportation systems (ITSs). Advanced Traveler Information Systems (ATIS) is one of the emerging technologies that will help travelers plan routes and schedules of their trips so as to redistribute the traffic over the highway network. Such redistribution will try to maximize the use of available highway capacity. Collections of real-time data and short-term predictions of traffic volumes are among the critical needs of an ATIS. This article studies characteristics of different traffic volume time series. In particular, time-series analysis is applied to the prediction of daily traffic volumes. The daily traffic volume is estimated by using the previous 13 daily traffic volumes. The study involves a comparison of statistical and neural network techniques for time series analysis. The analysis is applied to different types of road groups according to the trip purpose and trip length distribution. It is hoped that this study will provide a better understanding of various issues involved in the short-term prediction of traffic volumes on different types of highways.

Neural networks as alternative to traditional factor approach of annual average daily traffic estimation from traffic counts

Presented in this paper is a comparison of the neural network approach and the traditional factor approach for estimating annual average daily traffic (AADT) from 48-h sample traffic counts. Minnesota’s automatic traffic recorder (ATR) sites are investigated. The traditional AADT estimation approach involves application of volume adjustment factors to sample counts. The neural network model used in this study is based on a multilayered, feed-forward, and back-propagation design for supervised learning. The results of AADT estimation from a single shortperiod traffic count indicate that as compared with the neural network approach, the estimation errors for the factor approach can be lower under a scenario in which ATR sites are grouped appropriately and the sample sites are correctly assigned to various ATR groups. Unfortunately, the current recommended practice offers little guidance on how to achieve the assignment accuracy that may be necessary for obtaining reliable AADT estimates from sample counts. The advantage of the neural network approach is that classification of ATR sites and sample site assignments are not required. The neural network approach can be particularly suitable for estimating AADT from two or more shortperiod traffic counts taken at different times during the counting season.

Statistical Investigations of Statutory Holiday Effects on Traffic Volumes

Traffic volume fluctuates from time to time and from location to location, with significant variations in demand as a result. The increases in travel during statutory holiday periods are substantial, and some critical traffic problems have been reported. An understanding of this substantial variation in the volume of traffic can assist transportation agencies in developing practical countermeasures in aspects such as traffic control plans, signal timing, safety programs, traffic volume monitoring, and prediction. The literature on holiday traffic is limited, and no effort has been made to examine statistically the significance of changes in traffic volume due to holiday effects. With the past 20 years of data collected by permanent traffic counters on highways in Alberta, Canada, holiday effects on road traffic are shown graphically. Then, the nonparametric Wilcoxon matched pair test is used to test the variation characteristics of normal flow, the Friedman method is applied to investigate the holiday eff...

ESTIMATION OF ANNUAL AVERAGE DAILY TRAFFIC ON LOW-VOLUME ROADS: FACTOR APPROACH VERSUS NEURAL NETWORKS

Estimation of the annual average daily traffic (AADT) for low-volume roads is investigated. Artificial neural networks are compared with the traditional factor approach for estimating AADT from short-period traffic counts. Fifty-five automatic traffic recorder (ATR) sites located on low-volume rural roads in Alberta, Canada, are used as study samples. The results of this study indicate that, when a single 48-h count is used for AADT estimation, the factor approach can yield better results than the neural networks if the ATR sites are grouped appropriately and the sample sites are correctly assigned to various ATR groups. Unfortunately, the current recommended practice offers little guidance on how to achieve the assignment accuracy that may be necessary to obtain reliable AADT estimates from a single 48-h count. The neural network approach can be particularly suitable for estimating AADT from two 48-h counts taken at different times during the counting season. In fact, the 95th percentile error values of about 25 percent as obtained in this study for the neural network models compare favorably with the values reported in the literature for low-volume roads using the traditional factor approach. The advantage of the neural network approach is that classification of ATR sites and sample site assignments to ATR groups are not required. The analysis of various groups of low-volume roads presented also leads to a conclusion that, when defining low-volume roads from a traffic monitoring point of view, it is not likely to matter much whether the AADT on the facility is less than 500 vehicles, less than 750 vehicles, or less than 1,000 vehicles.

Imputation of Missing Traffic Data during Holiday Periods

Abstract Highway and transportation agencies implement large-scale traffic monitoring programs to fulfill the planning, operation and management needs of highway systems. These monitoring programs typically use inductive loops as detectors to collect traffic data. Because of the harsh environment in which they operate, they are highly prone to malfunctioning and providing erroneous or missing data. If this occurs during holiday periods when the increase in highway traffic is often substantial, there is a good chance that traffic peaking and variation will be underestimated. This paper discusses the adaptability of available imputation techniques for holiday traffic and then introduces a new procedure using non-parametric regression – the k-nearest neighbor (k-NN) method. It is found that the performance of the k-NN method is consistent and reasonable for different holidays and types of highway. In addition, it is also concluded that the data requirements for this method are flexible.

Refining Genetically Designed Models for Improved Traffic Prediction on Rural Roads

Research into advanced traveler information systems (ATIS) for rural roads is limited. However, highway agencies expect to implement intelligent transportation systems (ITS) in both urban and rural areas. In this paper, genetic algorithms (GAs) are used to design both time delay neural network (TDNN) models as well as locally weighted regression (LWR) models to predict short-term traffic for two rural roads in Alberta, Canada. A top-down refinement was used to study the interactions between modeling techniques and underlying data sets for obtaining highly accurate models. It is found that LWR models achieve faster accuracy improvement than TDNN models over the refinement process. Compared with previous research, the models proposed here show higher accuracy. The average errors for the best LWR models obtained through the model-refining process are less than 2% in most cases. For refined TDNN models, the average errors are usually less than 6–7%. The resulting models indicate a level of high robustness over different types of roads, and thus may be considered desirable for real-world statewide ITS implementations.

Variation of Impact of Cold Temperature and Snowfall and Their Interaction on Traffic Volume

This paper presents a detailed investigation of highway traffic variations with severity of cold, the amount of snow, and various combinations of cold and snow intensities. Separate analyses for starting, middle, and ending months of winter seasons are conducted to study the variations in traffic-weather relationships within the winter season. The study is based on hourly traffic flow data from 350 permanent traffic counter sites located on the provincial highway system of Alberta, Canada, and weather data obtained from nearby Environment Canada weather stations, from 1995 to 2005. Multiple regression analysis is used in the modeling process. The model parameters include three sets of variables: the amount of snowfall as a quantitative variable, categorized cold as a dummy variable, and an interaction variable formed by the product of these two variables. The study results indicate that the association of highway traffic flow with cold and snow varies with day of week, hour of day, and severity of weather conditions. A reduction of 1% to 2% in traffic volume for each centimeter of snowfall is observed when the mean temperature is above 0°C. For the days with zero precipitation, reductions in traffic volume due to mild and severe cold are 1% and 31%, respectively. An additional reduction of 0.5% to 3% per centimeter of snowfall results when snowfall occurs during severe cold conditions. Study results show lesser impact of adverse weather conditions on traffic during severe winter months (mid-November to mid-March) and the months thereafter compared with early (starting) winter months.

Assessing Robustness of Imputation Models Based on Data from Different Jurisdictions: Examples of Alberta and Saskatchewan, Canada

The literature indicates that many highway and transportation agencies in North America and Europe estimate missing values in their collected traffic data records. Estimating missing values is known as data imputation. Such a convention can be traced back to early traffic monitoring systems in the 1930s; however, no studies have been found to assess the accuracy of imputations carried out by transportation practitioners. The imputation methods used by highway agencies are varied and intuitive in nature. Some of them could result in large imputation errors in certain circumstances. Those errors can lead to significant deviations in the resulting operation plans and designed structures. Therefore, it is necessary to evaluate the accuracy of various imputation methods that highway agencies use. This study identifies and tests typical imputation methods on automatic traffic recorder data from Alberta and Saskatchewan in Canada. With assessment of imputation methods based on the data from different highway agencies, it is possible to evaluate their robustness and suitability for use across jurisdictions. The accuracy of individual imputation models was statistically analyzed, and comparisons and recommendations were made. Study results clearly indicate that models using additional observations as input and more sophisticated prediction techniques consistently produce better imputations. It is believed that this study would be helpful for traffic engineers in reviewing their imputation practices and, hence, in improving their data quality.

Matching Patterns for Updating Missing Values of Traffic Counts

Abstract The presence of missing values is an important issue for traffic data programs. Previous studies indicate that a large percentage of permanent traffic counts (PTCs) from highway agencies have missing hourly volumes. These missing values make data analysis and usage difficult. A literature review of imputation practice and previous research reveals that simple factor and time series analysis models have been applied to estimate missing values for transport related data. However, no detailed statistical results are available for assessing imputation accuracy. In this study, typical traditional imputation models identified from practice and previous research are evaluated statistically based on data from an automatic traffic recorder (ATR) in Alberta, Canada. A new method based on a pattern matching technique is then proposed for estimating missing values. Study results show that the proposed models have superior levels of performance over traditional imputation models.

Evolutionary Regression and Neural Imputations of Missing Values

While the information age has made a large amount of data available for improved industrial process planning, occasional failures lead to missing data. The missing data may make it difficult to apply analytical models. Data imputation techniques help us fill the missing data with a reasonable prediction of what the missing values would have been.